Re: [rtcweb] Current state of signaling discussion

[Hadriel Kaplan <[e-mail]>]

On Oct 18, 2011, at 5:18 PM, Eric Rescorla wrote:

> One might already have much of this stuff in your application to support
> whatever person-to-person communications (E.g., IM) you already
> have. When I heard "20 lines of code" I always assumed that it excluded that
> sort of routing-type stuff.

But that's the point exactly - OF COURSE that type of stuff might already be in your application, and could be conveyed in a URL or all be implicit or whatever.  Or it might not at all be in your app and need to added for this in some way.  

And that's the point many of us have been making: don't define a standard signaling protocol, because it's all application-specific to the web-app, which only the web-app developer would possibly know.

Thus, I'm asking the chairs what they mean when they use the word "signaling".  If they mean session signaling, then ROAP doesn't do it either.  If they only mean media-setup+negotiation signaling, that's a different problem scope. (I think that too is unnecessary, but if you put full SDP offer/answer state in the browser, then it's probably a fait accompli)

-hadriel